Machine Learning Scientist

About Hytech

Hytech is a premier management consulting firm based in Australia and Singapore, specializing in enabling digital transformation within fintech and financial services sectors. We offer comprehensive consulting services alongside reliable middle- and back-office solutions that help our clients streamline operations, boost efficiency, and maintain a competitive edge in a rapidly evolving digital environment. Our clientele features prominent global trading platforms and leading cryptocurrency exchanges.

We leverage AI and data-driven methodologies to address practical financial challenges such as risk management, operational optimization, and decision support, prioritizing impactful real-world outcomes.

With a global workforce exceeding 2,000 professionals and offices spanning Australia, Singapore, Malaysia, Taiwan, the Philippines, Thailand, Morocco, Cyprus, Dubai, and more, Hytech solidifies its strong international presence.

Role Overview

Join our quantitative trading division as a Machine Learning Scientist, tasked with designing, researching, prototyping, and deploying ML models tailored for crypto derivative market-making and algorithmic trading strategies. This role demands close collaboration with quantitative researchers, traders, and low-latency engineers to transform statistical and machine learning insights into operational trading platforms. Key focus areas include applied machine learning, time-series analysis, market microstructure research, and backtesting aimed at uncovering alpha, enhancing quoting, optimizing risk control, and improving execution quality across multiple centralized crypto exchanges.

Key Responsibilities

• Research and develop machine learning models for understanding market microstructure, order book behavior, alpha signal generation, price forecasting, spread optimization, and execution cost evaluation in crypto derivatives.
• Establish robust backtesting frameworks incorporating walk-forward validation and out-of-sample tests to assess model robustness against overfitting, market regime shifts, and exchange-specific anomalies.
• Create and engineer features from high-frequency datasets including order book snapshots, tick data, trade flows, funding rates, volume imbalances, queue dynamics, and latency statistics.
• Collaborate with traders and quantitative researchers to transform trading hypotheses into scalable ML model architectures and iteratively enhance models based on live trading feedback and performance indicators.
• Integrate trained ML models into low-latency market-making and execution systems, partnering with systems engineers to maintain a balance between model accuracy and latency requirements.
• Design monitoring tools and dashboards to observe live model performance indicators such as feature drift, concept drift, prediction accuracy, and profit and loss metrics.
• Conduct statistical investigations into market regimes, liquidity dynamics, limit order behaviors, and trading interruptions to refine model assumptions and risk management parameters.
• Produce thorough documentation covering research outputs, model specifications, testing methodologies, and deployment conditions for internal knowledge-sharing and compliance purposes.
• Support stress testing and scenario-based simulations to evaluate model resilience under volatile conditions, rate limiting, and exchange disruptions.

Candidate Requirements

• Graduate degree (Master’s or PhD) in Computer Science, Statistics, Applied Mathematics, Physics, Quantitative Finance, or related quantitative disciplines.
• Minimum of three years hands-on experience developing production-grade machine learning models within systematic trading, market-making, or high-frequency trading domains, preferably in crypto or traditional financial markets.
• Strong expertise in time-series machine learning techniques including gradient boosting, neural networks, sequence modeling, probabilistic methods, anomaly detection, and feature engineering specialized for high-frequency data.
• Proficient in Python with experience using NumPy, Pandas, Scikit-learn, PyTorch or TensorFlow, along with sound understanding of statistical principles such as hypothesis testing, probability theory, Bayesian inference, and correlation analysis.
• Comprehensive knowledge of market microstructure encompassing order types, limit order books, funding mechanisms, liquidity concepts, rate limits, trading halts, and exchange API behaviors.
• Experience implementing and maintaining backtesting frameworks with capabilities for performance attribution and managing overfitting challenges within financial time series data.
• Familiarity with real-time data streaming technologies and event-driven architectures using Kafka, PubSub, or Redis streams for feature extraction and ingestion.
• Exposure to low-latency trading environments is highly desirable; capability to collaborate with engineers to deploy models written in low-level languages like C++ or Rust is a plus.
• Preferable knowledge of crypto derivatives such as perpetual futures, centralized exchange market dynamics, and exchange-specific peculiarities.
• Demonstrated ability to maintain high coding standards, apply rigorous experimental validation, and effectively communicate sophisticated quantitative analyses to stakeholders without machine learning backgrounds such as traders and engineers.

Additional Preferred Qualifications

• Hands-on experience with online learning methodologies, reinforcement learning focused on market-making or order execution.
• Familiarity with performance profiling, tuning for low-latency systems and deploying containerized applications using Docker within cloud ecosystems.
• Contributions to academic publications or open-source projects related to time-series forecasting or quantitative finance are advantageous.
• Experience developing machine learning-driven risk models for real-time position monitoring.